The ultrashort light pulse compression technique, which allows realizing a minimum time duration that humankind can ever achieve, is yielding new discoveries and new technologies in so far unknown spheres of science with the use of ultrashort light pulses. Also, the ultrashort light pulse phase compensation technique with the ability to provide a train of light pulses with a controlled amount of phase compensation that varies with time commences being applied to various types of modulation spectroscopy and optical communication. Further, the ultrashort light pulse waveform shaping technique, which permits a light pulse to be shaped in any waveform as desired, is becoming indispensable in the elucidation of an elementary chemical reaction process and a biological reaction mechanism.
Heretofore, the ultrashort light pulse compression, the ultrashort light pulse phase compensation and the ultrashort light pulse waveform shaping have been attained upon phase-compensating each of spectral components of a light pulse generated by a pulsed light source and then adding together these phase-compensated spectral components.
For phase compensation, there are methods using a stationary optical element such as a prism pair, a grating pair or a dielectric multi-layer film mirror. However, since the phase compensation by these methods is stationary and, if dynamic, is not autonomous, it can be only effective where the phases of the spectral components of a light pulse are known beforehand and also temporally constant.
There is also a method for phase compensation by means of a 4-f pulse shaper using a SLM (space light modulator or spatial light modulator). See, for example, JP 2002-131710 A published May 9, 2002 on a patent application filed in the same inventorship as the present patent application. The method allows the amount of phase modulation to be varied dynamically and an optimum amount of phase modulation to be determined on a trial and error basis by using a simulated annealing or a genetic algorithm technique. However, since there need be a considerably large number of light pulse trains and a plenty of time before a phase determination is made, the method is found only efficient where the phases of the spectral components of a light pulse are temporally constant.
Thus, a problem in the prior art apparatus is that it has been necessary to use a light source only after determination is made in detail of phase information of a light pulse that the light source generates. Also, to reduce temporal fluctuations in phase of individual spectral components of a light pulse, its light source must have been equipped with a control system of extremely high grade. Further, such temporal fluctuations in phase if reduced by using a light source with a high degree of controllability have left the problem unresolved that the continuation time in which the ultrashort light pulse compression, phase compensation and waveform shaping can continuously be performed and accomplished is short, because of the unavoidable long-term fluctuation in phase.
In an effort to solve these problems, a method has recently been tried wherein a portion of output light pulses are taken out and these phases are continuously measured to feed the results back to a spatial light modulator so that a fluctuation in phase of the ultrashort light pulses if brought about is quickly compensated for autonomously (See, for example, JP 2002-131710 A above of the present inventors).
In this autonomous phase compensation method, phases are measured using an autocorrelation, a FROG (frequency resolved optical grating) or a SPIDER (Spectral Phase Interferometry for Direct Electric field Reconstruction) technique.
Of these techniques, however, the autocorrelation and FROG techniques have presented the problem that this method is so time-consuming that a usual laser light source whose temporal fluctuations in phase are large generally tend to fluctuate in phase during the measurement.
In the SPIDER method, two replica light pulses of an output light pulse are shared from the output light pulse and the two replica light pulses are delayed one after the other. The two light pulses delayed one after the other and a chirped light pulse shared from the output light pulse are introduced into a nonlinear optical crystal wherein the two replica light pulses have frequencies mixed with different frequencies in the chirped light pulse. The resultant frequency-mixed two replica light pulses are interfered with each other in a spectrometer to form an interference figure, from whose fringe spacing phase information of the output light pulse is extracted. See literature: IEEE Journal of Quantum Electronics, Vol. 35, No. 4, April 1999, p. 501–509. This method is capable of extracting all the spectral phase information at once of a single output light pulse and hence allows phase compensation at high speed. However, the need to form a chirped light pulse upon splitting the single, common output light pulse comes to diminish the intensity of replica light pulses and in turn to reduce the sensitivity at which the phase information can be extracted. The problem has thus been met that the SPIDER method as it is conventional is only effective for an output light pulse that is rather exceedingly high in intensity.
With the conventional ultrashort light pulse compression, phase compensation and waveform shaping apparatus, it is seen, therefore, that where a usual laser light source is used that is low in output light intensity or large in phase fluctuation, it has been impossible to autonomously perform the compression, phase-compensation or waveform-shaping of an ultrashort light pulse.